This invention relates to radiometric scatter measuring.
In a typical optical lithographic (also called photolithographic) process, as illustrated in FIG. 1, within an optical lithography tool 500 an image is copied from a reticle (mask) 100 onto a wafer 200 by projecting light 350 from a light source 300 through an illumination slot (frame) 150 over the patterned reticle 100, through precision optics (lenses) 400, onto the wafer 200. The wafer 200 has first been coated with a photoresist material (also called resist) that responds to light exposure in a specific, predictable way, e.g., by being made more or less soluble to a developer liquid upon illumination by light of a particular wavelength for a certain duration, under proper conditions.
In the ideal photolithographic process illustrated by FIG. 1, the imaged light 350 creates only an image on the wafer 200 that perfectly corresponds to the image or pattern of the reticle 100. In real-world photolithographic processes, however, as illustrated by FIG. 2, scattered light 370 (sometimes also known as flare or stray light) can, among other things, reduce the potential precision or efficiency of the system by reducing image contrast and process tolerances. The level of scatter can vary from system to system, within a single system, and over time.
The lithographic measurement of scattered light is commonly expressed as the percentage of open frame light intensity (i.e., the intensity of light near the surface of wafer 200 when reticle 100 is omitted or is transparent) measured in shadowed areas. One lithographic method of measuring scatter, described by Kirk, involves gradually increasing the light intensity (dose) over an opaque reticle feature until the image of the opaque feature formed in positive photoresist becomes fully developer-soluble. In other words, the opaque feature""s image (which would not develop away in an ideal system) is subjected to scattered light to the point that it does develop away.
One opaque feature commonly used with this method is a 2 xcexcm-wide isolated line in a clear field. This method requires observation of the dose in the clear surrounding area when the photoresist there develops away, followed by observation of the dose step when the image of the opaque line develops away. The scatter measurement is commonly expressed as a percentage, dividing the dose to clear the surrounding resist (Eo) by the dose to clear the line (Do). The larger the line""s width, the further the light has to scatter to clear all the resist.
Measuring scatter as a function of feature size in this way has been a popular method for determining the scatter distributionxe2x80x94also called the point spread function (PSF)xe2x80x94of a photolithography system at relatively short ranges of spread (i.e., scatter at relatively short distances in the x-y plane from the light source). A second, similar lithographic scatter measurement method also described by Kirk is based on observing the receding edge of a photoresist image of an opaque reticle feature as the dose is increased and the edge recedes into the dark region of the image.
These lithographic scatter measurement methods already in use involve the measurement of Do along the length of the illumination slot, to track the changes in value. These methods can be tedious, time-consuming, and resource-intensive. Moreover, potential sources of error or imprecision in such measurement methods include the variation of Eo within a wafer and from wafer to wafer, the discrete dose increments used during testing, and human variability and error in the visual inspection steps of the process. In addition, photoresist-generated lens contamination can change the level and shape of the scatter PSF over time, and these effects can be overlooked by the standard lithographic measurement methods that use only one feature size, such as a 2 xcexcm isolated line.